Scott Davis, PhD., P.E., CFEI
Tom DeBold, P.E., CFEI
and
John Pagliaro, Ph.D.
Gexcon US, USA
Presented at the International Symposium on Fire Investigation Science and Technology, 2018
ABSTRACT
When investigating a flash fire or explosion, it is common to test and evaluate the integrity of a fuel gas system – such as LPG systems. One common technique is to use a pressurized inert gas, such as nitrogen, to not only identify possible holes for propane to leak through but also quantify the flow rate of the leak. However, if the leak is located under the liquid level of the propane cylinder, there is a common misconception that the mass flow rate for a liquid propane release must be approximately 270 times larger than a vapor release through the same hole size. In a recent case, an investigator concluded that a small hole identified during a leak test using nitrogen gas would be more than sufficient to cause a large flash fire because the hole would have been below the liquid propane level and thus would be 270 times larger than a vapor propane leak through the same hole. This misconception regarding the magnitude of liquid propane leaks stemmed from the fact that liquid propane has a density that is approximately 270 times larger than the density of gaseous propane at atmospheric pressure and temperature. Under practical conditions, the mass flow rates of liquid propane will actually be much closer to those of vapor releases given identical hole sizes. In fact, the difference will be only approximately 6-8 times larger if both liquid and gaseous propane are stored at ambient temperatures of 20°C. This near 30-fold reduction in liquid flow is due to how propane is stored and the physics governing a compressible vapor release compared to an incompressible liquid release. More specifically, for propane at 20°C the vapor will exit a small hole at the speed of sound and due to the high storage pressure (e.g., approximately 105 psig at 20°C) the density of the vapor leaking will be much higher than atmospheric conditions. Conversely, an incompressible liquid will be limited by the pressure difference upstream and downstream of the hole. The present study demonstrates the differences in liquid and vapor propane release rates and shows that liquid releases are not 270 larger than vapor releases. This is accomplished by: (1) providing an overview of the theoretical orifice flow equations for pressurized vapor and liquid releases; (2) presenting results from actual releases of liquid and vapor propane from a 120-gallon water capacity tank and a 20 lb cylinder; and (3) presenting CFD results that highlight the importance of accurate estimates of leak rates when it comes to formulating origin and cause hypotheses.
Download Complete Paper Here